CN102130746B - Network coding feedback method for coordinated multi-point transmission system - Google Patents

Abstract

The invention discloses a network coding feedback method in a coordinated multi-point transmission system, which belongs to the technical field of communication. The information that needs to be fed back (such as PMI information) and the information that is fixed to be sent (such as RS information) are first network coded and then sent out. The idea of network coding is introduced to transmit feedback information at the same time while keeping the existing uplink transmission overhead unchanged. Since the transmitted information sequence after network encoding still has good cross-correlation and autocorrelation, the PMI information is obtained by correlation detection and searching for the minimum Euclidean distance at the receiving end. The present invention takes the SRS information that the system needs to transmit as the carrier, The information that needs to be fed back is hidden in the network-coded SRS sequence through network coding, and the transmission efficiency is improved by reducing or even eliminating the feedback overhead, effectively improving the system throughput.

Description

Network Coding Feedback Method in Coordinated Multipoint Transmission System

technical field

The invention relates to a feedback transmission scheme in a communication system, in particular to a network coding feedback method for reducing system feedback overhead on the basis of maintaining system transmission reliability, and belongs to the technical field of communication. the

Background technique

In the coordinated multi-point transmission (CoMP) system, the design of the feedback scheme has a great impact on the system performance. Various implicit feedback, explicit feedback and feedback schemes based on SRS (Sounding RS) are currently hot topics discussed by major companies. . In the design process of various feedback schemes, feedback overhead is a very important factor to consider. It is necessary to keep the degradation of the performance of the feedback scheme as little as possible under the premise of reducing the feedback overhead as much as possible. For example, in the implicit feedback scheme, the feedback information is generally limited to A precoding vector index or a codebook index with a limited number of bits. the

But no matter how the amount of feedback is compressed, the feedback information needs to be transmitted separately, which will undoubtedly increase the additional feedback burden. So can the feedback information be transmitted at the same time while keeping the existing uplink transmission overhead unchanged? the

Feedback based on network coding can solve this problem. In the feedback process, the user terminal (UE) performs network coding on the information to be fed back (such as PMI information) and the information (such as RS information) to be sent by itself, and sends it out. After decoding, the receiving end can separate the original sent information and Feedback information. the

The LTE physical layer defines two types of uplink RS:

(1) Demodulation RS (DMRS: Demodulation RS), mainly used for coherent demodulation in channel estimation;

(2) Sounding RS (SRS: Sounding RS), which is not associated with uplink data and control transmission, is used to determine channel quality, thereby performing frequency selective scheduling in the uplink. the

At present, sequences with Constant Amplitude Zero Auto-Correlation (CAZAC) characteristics are mostly used as uplink reference signal (DMRS/SRS) sequences. DMRS is sent together with data or control information, and its resource location (time and Frequency) is determined by the corresponding shared channel or control channel. The transmission of the uplink Sounding reference signal has nothing to do with the uplink physical channel, and is an independent uplink channel. According to a predefined period, the terminal transmits at the frequency domain position where channel measurement needs to be performed. The network coding discussed in the present invention refers to that the UE end performs network coding with PMI and SRS to obtain the sequence SRS ePMI', and feeds it back to the eNB (base station) side; after receiving it, the eNB compares the network coding under all PMI situations Get the SRS ePMI', and get the original PMI information. the

Contents of the invention

Aiming at the problem that the feedback overhead of the existing feedback scheme in the CoMP system is large, the present invention proposes a network coding feedback method that can greatly reduce (or even completely eliminate) the feedback overhead without losing system reliability. In order to improve the system effectiveness and reduce the feedback overhead, the present invention introduces the idea of network coding, and performs modulo 2 exclusive OR operation on the linearly coded PMI' sequence and the SRS sequence at the UE. the

The base stations eNB1 and eNB2 serve the user UE at the same time. After a downlink transmission, the UE not only needs to send the pilot information (that is, the SRS sequence) of the eNB1, but also needs to feed back the PMI information of the eNB2 link to the eNB1. It is known that in the SRS sequence set to be transmitted by the UE, there are N kinds of SRS sequences generated by different root indices, which are completely orthogonal to each other. the

In order to achieve the above object, the network coding feedback method in the coordinated multi-point transmission system described in the technical solution of the present invention comprises the following steps:

Step 1, perform linear block coding on the 2 ^k PMI information of the eNB2 link of the user UE that may need to be fed back to the eNB1, that is, 2 ^k k-bit PMI _i (i=0, 1, .., 2 ^k -1 ) is linearly block coded, mapped to 2 ^k long sequences of length L bits, denoted as PMI'_i; L is the length of the SRS sequence;

For example, when k=4, there are 16 PMI _i in total, and i takes a value from 0 to 15; each user UE must transmit one of these 16 PMI _i , k is preset according to system performance requirements, k The longer it is, the more information PMI _i carries;

Step 2, perform network coding on each PMI' _i obtained in step 1 and each SRS sequence in the SRS sequence set to be transmitted by the user UE according to the following formula:

${\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; NC</span>}}^{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; SRSe</span>}{\mathrm{<span\; class="notranslate">\; PMI</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

The SRS sequence set includes N SRS sequences, the length of each SRS sequence is L, that is, equal to the length of _PMI'i , each transmission only needs to transmit one SRS sequence, and the length L is determined by the actual system requirements;

作为网络编码的码字集合S，并建立一个相应的码表，用于存储构成该集合中每个网络编码码字 

所对应的SRS和PMI_i；  The obtained N×2 ^k network coding codewords

As the codeword set S of network coding, and establish a corresponding code table for storing each network coding codeword in the set

Corresponding SRS and PMI _i ;

The network coding operation ⊙ described here is to perform an XOR operation on two binary sequences to obtain a new "01" sequence after network coding. the

Preferably, the SRS sequence described here is the pilot information of eNB1 to be sent by the user UE. the

In the SRS sequence set to be transmitted by the user UE, there are N kinds of SRS sequences generated by different root indices, which are completely orthogonal to each other. the

并将这个 

经过调制后向eNB1发送出去，eNB1接收并存储为序列 

其中，q为0到2^k-1中任意一个整数(含边界)，即与i取值范围相同；这种情况在通信双方可以保持同步(即发射端也能调用步骤2建立的码表)时可以节省很大的计算量，不用每次发射都进行编码和异或运算，直接查表就可以得到对应的网络编码码字 

Step 3. According to the current need to feed back the PMI _q of the eNB2 link to eNB1 and the current SRS sequence to be transmitted, the user UE searches for the corresponding network coding codeword from the code table established in step 2

and put this

After modulation, it is sent to eNB1, and eNB1 receives and stores it as a sequence

Among them, q is any integer from 0 to 2 ^k -1 (including the boundary), that is, the value range of i is the same; in this case, the two sides of the communication can maintain synchronization (that is, the transmitter can also call the code table established in step 2) It can save a lot of calculation, instead of encoding and XOR operation for each transmission, you can directly look up the table to get the corresponding network encoding codeword

并将这个 

经过调制后向eNB1发送出去，eNB1接收并存储为序列 

这种情况适用于通信双方无法可以保持同步(即发射端不能调用步骤2建立的码表)的系统。  Or perform linear block coding on the PMI _i at the UE to obtain the _PMI'q sequence, and then perform network coding on the _PMI'q and the current SRS sequence to be transmitted as described in step 2 to generate

and put this

After modulation, it is sent to eNB1, and eNB1 receives and stores it as a sequence

This situation is applicable to the system where the communication parties cannot maintain synchronization (that is, the transmitter cannot call the code table established in step 2).

进行比较，即按照如下原则估计实际发送的序列：  Step 4, eNB1 searches in the N×2 ^k network coding codewords of the network coding codeword set S, and the received codewords

For comparison, that is, to estimate the actual sent sequence according to the following principles:

$\stackrel{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; S</span>}}{\mathrm{<span\; class="notranslate">\; RS</span>}}_{\mathrm{<span\; class="notranslate">\; NC</span>}}^{\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; arg</span>}\underset{{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; NC</span>}}^{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; S</span>}}{\mathrm{<span\; class="notranslate">\; min</span>}}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; NC</span>}}^{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; NC</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}_{\mathrm{<span\; class="notranslate">\; f</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1,2</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

为Frobenius范数的平方—这里是把 和 

当作向量来处理的。上面公式实际就是要在集合S中这N×2^k个可能的网络编码码字中寻找一个与 

序列欧式距离最近的网络编码，即两者作差运算后得到的序列向量模值最小。其中 

唯一对应集合S中N×2^k个网络编码码字的一个，找到这个使范数平方最小的网络编码就意味着同时找到了用户UE实际发送的PMI_q序列和SRS序列；  In the formula, S is the code word set after the network coding established in step 2,

is the square of the Frobenius norm—here the and

treated as a vector. The above formula is actually to find one of the N×2 ^k possible network coding codewords in the set S that matches

The network coding with the closest sequence Euclidean distance, that is, the sequence vector modulus obtained after the difference operation between the two is the smallest. in

It uniquely corresponds to one of the N×2 ^k network codewords in the set S, and finding this network code that minimizes the norm square means that the PMI _q sequence and the SRS sequence actually sent by the user UE are found at the same time;

表示UE实际发送的网络编码序列，步骤4中的 

表示接收端估计的UE发送的网络编码序列，在没有差错的情况下两者是一样的，有差错情况下有可能误判为另一个网络编码。  Note that in step 3 the

Indicates the network coding sequence actually sent by the UE, in step 4

Indicates the network code sequence sent by the UE estimated by the receiving end. If there is no error, the two are the same. If there is an error, it may be misjudged as another network code.

以后，根据步骤2建立的码表(即所有可能的SRS序列和所有可能的PMI′_i序列异或运算得到的网络编码码字表)找到对应的SRS和 PMI′_q。由于事先对PMI_i做的是线性分组编码，所以找到码表中的码字PMI′_q以后只需要取分组码字PMI′_q的前k位即为PMI_q。  Step 5, OK

Afterwards, find the corresponding SRS and PMI' _q according to the code table established in step 2 (that is, the network coding code word table obtained by XOR operation of all possible SRS sequences and all possible PMI' _i sequences). Since the PMI _i is linearly block coded in advance, after finding the code word PMI′ _q in the code table, only the first k bits of the block code word PMI′ _q need to be obtained as PMI _q .

应具有良好的自相关性和互相关性，当保证网络编码码字仍具有较好的自相关性和互相关性时，可以利用码字集合S中的 

按照传统方法在受到干扰的接收信号，即下式中的信号Y(小区发射信号时会对周围邻小区接收信号形成干扰)中提取出匹配基站的信号进行分析，得到信道状态信息。下面从理论上说明本发明提出的网络编码反馈方法具有实际可操作性，能够用来区分不同用户的信号。  As a preferred solution, in step 1, the k-bit PMI information to be linearly block-coded is determined according to the information that the actual system needs to transmit, but how to map (ie, linearly encode) the k-bit PMI sequence to the L-bit _PMI'i Sequences have different mapping encoding rules. The present invention obtains the best mapping encoding rule for mapping and coding each k-bit PMI sequence of 2 ^k k-bit PMI sequences into L-bit PMI' _i sequences through computer search. The principle of searching is to follow this mapping coding rule, so that the k-bit PMI information sequence to be transmitted is subjected to the linear block coding described in step 1, and then the codeword obtained by network coding with SRS

It should have good autocorrelation and cross-correlation. When the network coding codewords still have good autocorrelation and cross-correlation, you can use the codeword set S

According to the traditional method, the signal of the matching base station is extracted from the interfered received signal, that is, the signal Y in the following formula (when the cell transmits the signal, it will interfere with the received signal of the surrounding adjacent cells) and analyzed to obtain the channel state information. The following theoretically demonstrates that the network coding feedback method proposed by the present invention is practical and operable, and can be used to distinguish signals of different users.

即PMI′_me SRS_j，表示第j个小区的识别序列，其中PMI′_m为需要反馈的PMI序列进行线形分组编码后的序列，m∈{0，1，2，...，2^k-1}，SRS_j为未进行网络编码前的完全正交的SRS序列，j∈{1，2，...，n}；不同 序列之间具有良好的自相关性和互相关性(如附图5所示)，利用这种性质为不同的基站找到自己的匹配信号。  Generally, at the receiving end of a certain cell, the signal will be interfered by other different cells. Use Y to indicate the received signal. There are n cells in total. X _j represents the signal transmitted by the jth cell. The network code word

That is, PMI′ _m e SRS _j represents the identification sequence of the jth cell, where PMI′ _m is the sequence of the PMI sequence that needs to be fed back and undergoes linear block coding, m∈{0, 1, 2,..., 2 ^k -1}, SRS _j is a completely orthogonal SRS sequence before network coding, j ∈ {1, 2, ..., n}; different The sequences have good auto-correlation and cross-correlation (as shown in Fig. 5 ), and use this property to find their own matching signals for different base stations.

$\mathrm{<span\; class="notranslate">\; Y</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; +</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; +</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}$

At the j-th base station, in order to extract the signal X _j transmitted by the j-th cell, the auto-correlation and cross-correlation of the codeword are coded by the network, so it is only necessary to perform an XOR operation on the received signal Y, that is

$\mathrm{<span\; class="notranslate">\; Ye</span>}{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; +</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; +</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}^{\mathrm{<span\; class="notranslate">\; m</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}$

Compared with the prior art, the beneficial effect of the present invention is that, on the premise of maintaining the reliability of information transmission, the communication source improves the transmission efficiency by reducing or even eliminating the feedback overhead. The introduction of the network coding idea can transmit the feedback information at the same time while keeping the existing uplink transmission overhead unchanged. Since the transmitted information sequence after network coding still has good cross-correlation and autocorrelation, at the receiving end, the PMI information can be obtained by correlation detection and finding the minimum Euclidean distance, and the received noise-added information sequence can be further used in the future The channel quality estimation value is obtained by sending the information sequence with the non-noise-added signal source. The present invention takes the SRS information that the system originally needs to transmit as the carrier, and hides the information that needs to be fed back in the SRS sequence after the network coding by means of network coding, which can reduce or completely eliminate (the difference between the PMI sequence and the SRS sequence) length decision) the unavoidable feedback burden in traditional feedback schemes, thus effectively improving the system throughput. the

Description of drawings

Fig. 1 is the feedback model diagram based on network coding that the present invention selects;

Fig. 2 is the CoMP system network coding feedback scheme designed by the present invention;

Fig. 3 is that the present invention has the PMI ' _i codebook figure of better cross-correlation after the coding that searched by computer;

Fig. 4 is the PMI decoding error rate figure obtained under the scheme of the present invention;

序列的相关性曲线图。  Fig. 5 is the 36 bit length after the network coding that the present invention selects

Correlation plot of the series.

Detailed ways

In order to make the object, technical solution and advantages of the invention clearer, the embodiments of the invention will be described in detail below in conjunction with the accompanying drawings. This embodiment is carried out on the premise of the technical solution of the invention, and the detailed implementation and specific operation process are given, but the protection scope of the present invention is not limited to the following embodiments. the

As shown in Figure 1, the base stations eNB1 and eNB2 serve the user UE at the same time (the present invention is limited to the case where two base stations serve one UE, and the two base stations serve the UE at the same time, regardless of sequence, and the two different transmissions only occupy One time slot can improve the system throughput, which is the advantage of the solution proposed in this patent). After a downlink transmission, the UE not only needs to send the pilot information (that is, the SRS sequence) of the eNB1, but also needs to feed back the PMI information of the eNB2 link to the eNB1. It is known that in the SRS sequence set to be transmitted by the UE, there are N kinds of SRS sequences generated by different root indices, which are completely orthogonal to each other. the

As mentioned above, the UE needs to transmit a k-bit PMI _i sequence and an SRS sequence each time. What the present invention needs to do is to find out the PMI _i and the corresponding SRS transmitted by the user UE; there are 2 ^k possible PMI _i sequences, By comparing the network coding results corresponding to all PMI _i sequences with the received sequences one by one, which one is most similar to which PMI _i is considered, that is, the transmitted PMI _i sequence is unknown to the receiving end, and the receiving end Only know all possible situations, and take one-to-one comparison to find the most likely result as the final solution.

In a communication system, in addition to transmitting the information required by the sink, the information source also needs to feed back its own unique information. In the existing explicit feedback, implicit feedback or SRS (Sounding RS)-based feedback schemes, no matter how the feedback volume is compressed, the feedback information needs to be transmitted separately, which will undoubtedly increase the additional feedback burden. The introduction of the network coding idea can transmit the feedback information at the same time while keeping the existing uplink transmission overhead unchanged. The specific method is to perform network coding on the information that needs to be fed back (such as PMI information) and the information that is fixed to be sent (such as RS information) before sending out. Since the transmitted information sequence after network coding still has good cross-correlation and autocorrelation, the PMI information can be obtained by correlation detection and finding the minimum Euclidean distance at the receiving end (as shown in Figure 2). the

The network coding discussed in the present invention refers to that the UE end performs network coding with PMI and SRS to obtain the sequence SRSe PMI', and feeds it back to the eNB (base station) side; after receiving it, the eNB compares the network coding under all PMI situations Get the SRS ePMI', and get the original PMI information. the

It is worth noting that since the length of the PMI information sequence to be transmitted is different from that of the SRS sequence, the relatively short PMI sequence must be mapped to a sequence PMI′ of the same length as the SRS sequence before the network coding operation of the two, so that after mapping The SRS' sequence obtained by network coding the PMI' and the SRS sequence has a good cross-correlation property. In order to extract the original PMI information simply at the receiving end, the present invention adopts the linear block code of the system to realize the one-to-one mapping from the PMI sequence to the PMI'. PMI' codes for cross-correlation properties. From the preliminary simulation results, it can be found that after network encoding of PMI information and SRS sequence, the feedback information PMI can be well restored at the receiving end. the

Without loss of generality, in this embodiment, it is assumed that the SRS length L=36, the PMI length k=4, and there are three possible SRS sequences, that is, N=3. the

Step 1, search out the best coding rule at the UE according to the computer, do linear block coding mapping on the PMI sequence with a length of 4, and obtain a PMI' _i sequence with a length of 36 bits, corresponding to 2 ⁴ =16 linear blocks of PMI Encoding code table shown in Figure 3 .

The best coding rule here is to make the code words obtained by network coding with SRS after linear block coding of PMI still have good auto-correlation and cross-correlation. the

In Fig. 3, Mapping has 16 rows and 36 columns in total, and each row is a codeword, which are linear block codewords corresponding to PMI ₀ = [0 0 0 0] to PMI ₁₅ = [1 1 1 1]. These codewords are obtained by computer search, and the code weight and the Hamming distance between the codewords are kept as equal as possible. In fact, the first 14 lines all have codewords with weight 18, and the last two codewords have weight 17, and these 16 linear block codewords have a Hamming distance between 16 and 20 from each other. According to the coding rules shown in Figure 3, the PMI sequence is linearly block-coded, and then the codewords obtained by network coding with SRS still have good autocorrelation and cross-correlation;

Step 2, the UE performs network coding operations on the SRS sequences with equal lengths of 36 bits and the _PMI'i sequences obtained in step 1,

${\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; NC</span>}}^{\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; SRS</span>}}^{\mathrm{<span\; class="notranslate">\; q</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; PMI</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; q</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; PMI</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; q</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; PMI</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}^{\mathrm{<span\; class="notranslate">\; q</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; PMI</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; 36</span>}}^{\mathrm{<span\; class="notranslate">\; q</span>}}\mathrm{<span\; class="notranslate">\; e</span>}{\mathrm{<span\; class="notranslate">\; PMI</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; 36</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>$

表示SRS^q中的第t位比特，t∈{1，2，...35，36}；PMI′_i表示16个可能的待发送的PMI线性编码后序列中的一个，i∈{0，1，...，15}，而PMI′_ij∈{0，1}表示PMI′_i序列中的第j位比特，j∈{1，2，...35，36}。  In the above formula, SRS ^q represents the SRS sequence to be sent with a length of 36 bits to be sent,

Represents the t-th bit in SRS ^q , t∈{1, 2,...35, 36}; PMI′ _i represents one of the 16 possible PMI linearly encoded sequences to be sent, i∈{0, 1, ..., 15}, and PMI′ _ij ∈ {0, 1} represents the jth bit in the PMI′ _i sequence, j ∈ {1, 2, ... 35, 36}.

作为网络编码的码字集合S，并建立一个相应的码表，用于存储构成该集合中每个网络编码码字 

所对应的SRS和PMI_i；  The obtained 3×2 ⁴ =48 network coding codewords

As the codeword set S of network coding, and establish a corresponding code table for storing each network coding codeword in the set

Corresponding SRS and PMI _i ;

经过调制后向eNB1发送出去，eNB1接收并存储为序列 

其中，q为0到15中任意一个整数(含边界)，即与i取值范围相同；  Step 3. According to the current need to feed back the PMI _q of the eNB2 link to eNB1 and the current SRS sequence to be transmitted, the user UE searches for the corresponding network coding codeword from the code table established in step 2 and put this

After modulation, it is sent to eNB1, and eNB1 receives and stores it as a sequence

Among them, q is any integer from 0 to 15 (including boundaries), that is, the value range is the same as that of i;

经过调制后向eNB1发送出去，eNB1接收并存储为序列  Or when the transmitter cannot call the code table established in step 2, perform linear block coding on PMI _i at the UE to obtain the PMI' _q sequence, and then perform the network operation as described in step 2 on the PMI' _q and the current SRS sequence to be transmitted encode, generate and put this

After modulation, it is sent to eNB1, and eNB1 receives and stores it as a sequence

有最小欧式距离的其中一个码字 认为是实际发送的序列：  Step 4, eNB1 searches for 3×2 ⁴ =48 network coding codewords obtained by XOR operation of all possible 3 kinds of SRS sequences and all possible 16 kinds of PMI′ _i sequences. Combine these 48 network encoding codewords with the received codewords compare to find

One of the codewords with the smallest Euclidean distance Considered to be the sequence actually sent:

$\stackrel{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; S</span>}}{\mathrm{<span\; class="notranslate">\; RS</span>}}_{\mathrm{<span\; class="notranslate">\; NC</span>}}^{\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; arg</span>}\underset{{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; NC</span>}}^{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; S</span>}}{\mathrm{<span\; class="notranslate">\; min</span>}}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; NC</span>}}^{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; SRS</span>}}_{\mathrm{<span\; class="notranslate">\; NC</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}_{\mathrm{<span\; class="notranslate">\; f</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0,1,2</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}$

Step 5, OK Afterwards, find the corresponding PMI _i according to the code table (that is, the network coding code word table obtained by XOR operation of all possible 3 kinds of SRS sequences and all possible 16 kinds of PMI _i sequences).

仍具有较好的互相关性和自相关性(如附图5所示)，所以可以利用得到的 

按照传统方法在受到干扰的信号(小区发射信号时会对周围邻小区接收信号形成干扰)中提取出匹配基站的信号进行分析，得到信道状态信息。  Since the linear block code searched by computer programming can guarantee the network coding

still has good cross-correlation and autocorrelation (as shown in Figure 5), so the obtained

According to the traditional method, the signal of the matching base station is extracted from the interfered signal (when the cell transmits the signal, it will interfere with the signal received by the surrounding adjacent cells) for analysis, and the channel state information is obtained.

加上高斯白噪声的干扰。PMI解码错误率是指解码错误的PMI个数在总的PMI中所占的比例；SNR是信号功率(SRS是恒福序列，分布在单位圆上，功率为1)对噪声功率的比值。从图4可以看出，本发明具有很好的传输可靠性。  According to the codewords in the above Mapping, the SRS is network coded and simulated. Figure 4 shows the effect of recovering the PMI at the receiving end after network coding. In the simulation model, only

Interference with Gaussian white noise. The PMI decoding error rate refers to the proportion of the number of decoding errors in the total PMI; SNR is the ratio of the signal power (SRS is a constant blessing sequence, distributed on the unit circle, and the power is 1) to the noise power. It can be seen from Fig. 4 that the present invention has good transmission reliability.

To ensure that the receiving end can distinguish and extract the corresponding signals, it is necessary to ensure good cross-correlation characteristics of the encoded sequences. The influence of network coding on the sequence correlation characteristics of SRS is studied separately below. the

和 

对应的根指数分别为q1＝25、q2＝29和q3＝34。序列 

跟 

和 

的相关特性分别记为R_1-1、R_1-2和R_1-2。网络编码使用序列随机产生，编码后的序列记为 

和  Take three SRS sequences

and

The corresponding root indices are q1=25, q2=29 and q3=34, respectively. sequence

and

and

Relevant properties of are denoted as R _1-1 , R _1-2 and R _1-2 , respectively. The network encoding is randomly generated using a sequence, and the encoded sequence is denoted as

and

When the sequence length L=36, the SRS sequence correlation characteristics before and after the network coding are shown in Fig. 5, and q=25 is drawn in the figure, the autocorrelation performance curve (as shown in Fig. Coordinates reach 1 shown in the band "o" type curve) and with all other possible 3 * 16-1=47 kinds of cross-correlation performance curves of different code words (band "o", "*", "x" type curve Respectively represent the sequence operation results corresponding to three different root indices q ₁ =25, q ₂ =29 and q ₃ =34).

与 

的互相关值一只维持在0.4以下，也就是说网络编码之后的不同SRS码字序列之间仍然可以通过相关特性加以区分。事实上，这里为原SRS做网络编码使用的是随机比特序列，如果能设计好线性分组码字，可获得更好的互相关特性。  As shown in Figure 5, before and after network coding, the autocorrelation curve of the SRS sequence has obvious impulse at 0 point,

and

The cross-correlation value of is only maintained below 0.4, which means that different SRS codeword sequences after network coding can still be distinguished by correlation characteristics. In fact, the network coding for the original SRS here uses a random bit sequence. If a linear block code can be designed, better cross-correlation characteristics can be obtained.

仍具有较好的互相关性和自相关性，所以可以按照传统方法在受到干扰的信号中提取出匹配基站的信号进行分析，得到信道状态信息。  Figure 5 shows that: after network coding

It still has good cross-correlation and autocorrelation, so the signal of the matching base station can be extracted from the interfered signal according to the traditional method for analysis, and the channel state information can be obtained.

The specific description above further elaborates the purpose, technical solution and beneficial effect of the invention. It should be understood that the above description is only a specific embodiment of the present invention and is not used to limit the protection of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention. the

Claims (8)

1. the network code feedback method in multipoint cooperative transmission system, base station eNB 1 and eNB2 for user UE service, is characterized in that simultaneously, comprise the following steps:

Step 1, by 2 of user UE ^kthe individual PMI information that may need to feed back to the eNB2 link of eNB1 is carried out Linear block coding, is about to 2 ^kthe PMI of individual k bit _icarry out Linear block coding, be mapped as 2 ^kthe long sequence of individual length L bit, is designated as PMI _i'; Wherein L is the length of SRS sequence, i=0, and 1 ..., 2 ^k-1; K is predefined according to systematic function requirement, and k is longer, PMI _ithe information of carrying is just more;

Step 2, each PMI according to the following formula step 1 being obtained _i' carry out network code with each the SRS sequence in the armed SRS arrangement set of user UE:

And PMI described SRS arrangement set comprises N SRS sequence, and the length of each SRS sequence is L, _i' equal in length, each transmission only needs to transmit a SRS sequence, length L is required to determine by real system; Network code computing ⊙ described herein carries out XOR to two binary sequences, obtains new " 01 " sequence after network code;

By N * 2 that obtain ^kindividual network code code word

as the codeword set S of network code, and set up a corresponding code table, for storing, form each network code code word of this set

corresponding SRS and PMI _i;

Step 3, user UE is according to the current PMI that need to feed back to the eNB2 link of eNB1 _q, and current armed SRS sequence, the code table of setting up from step 2, search corresponding network code code word

and by this through ovennodulation, backward eNB1 sends, and eNB1 receives and be stored as sequence SRS _nC; Wherein, q is 0 to 2 ^kany one integer in-1 is identical with i span;

Step 4, eNB1 is these N * 2 in the codeword set S of network code ^kin individual possible network code code word, find one with the code word SRS receiving _nCthe network code that sequence Euclidean distance is nearest, according to following principle, estimate the actual sequence sending:

${\mathrm{SRS}}_{\mathrm{NC}}^q=\arg \underset{{\mathrm{SRS}}_{\mathrm{NC}}^i\&Element;S}{\min }{\left|\right|{\mathrm{SRS}}_{\mathrm{NC}}^i-{\mathrm{SRS}}_{\mathrm{NC}}\left|\right|}_F^2,i=\mathrm{0,1,2},...,2^k-1---\left(2\right)$

In formula, S is the codeword set after the network code set up of step 2,

for Frobenius norm square; Wherein

n * 2 in unique corresponding S set ^ka network code of individual network code code word, makes norm squared minimum;

Step 5, determines after, the code table of setting up according to step 2 finds corresponding SRS and PMI ' _q; Get this grouping code word PMI ' _qfront k position be PMI _q.

2. the network code feedback method in a kind of multipoint cooperative transmission system according to claim 1, is characterized in that, described SRS sequence is the pilot frequency information of user UE eNB1 to be sent.

3. according to the network code feedback method in a kind of multipoint cooperative transmission system described in claim 1 or 2, it is characterized in that the SRS sequence that has N kind to be produced by different radical exponents in the armed SRS arrangement set of described user UE, each other complete quadrature.

4. according to the network code feedback method in a kind of multipoint cooperative transmission system described in claim 1 or 2, it is characterized in that, step 3 adopts following scheme: at UE place by PMI _icarry out Linear block coding and obtain PMI ' _qsequence, then by PMI ' _qcarry out the network code as described in step 2 with current armed SRS sequence, generate

and by this

through ovennodulation, backward eNB1 sends, and eNB1 receives and be stored as sequence SRS _nC.

5. the network code feedback method in a kind of multipoint cooperative transmission system according to claim 3, is characterized in that, step 3 adopts following scheme: at UE place by PMI _icarry out Linear block coding and obtain PMI ' _qsequence, then by PMI ' _qcarry out the network code as described in step 2 with current armed SRS sequence, generate

and by this

through ovennodulation, backward eNB1 sends, and eNB1 receives and be stored as sequence SRS _nC.

6. according to the network code feedback method in a kind of multipoint cooperative transmission system described in claim 1 or 2, it is characterized in that, in step 1, the k position PMI information of pending Linear block coding is that the information that need to transmit according to real system determines, by this 2 ^keach k position PMI sequence mapping of individual k position PMI sequence is encoded into the PMI of L position _ithe best mapping code rule of ' sequence is, makes to need after the k position PMI information sequence of the transmission Linear block coding described in carry out step 1, then carries out the resulting code word of network code with SRS

should there is good autocorrelation and cross correlation.

7. the network code feedback method in a kind of multipoint cooperative transmission system according to claim 3, is characterized in that, in step 1, the k position PMI information of pending Linear block coding is that the information that need to transmit according to real system determines, by this 2 ^keach k position PMI sequence mapping of individual k position PMI sequence is encoded into the PMI of L position _ithe best mapping code rule of ' sequence is, makes to need after the k position PMI information sequence of the transmission Linear block coding described in carry out step 1, then carries out the resulting code word of network code with SRS

should there is good autocorrelation and cross correlation.

8. the network code feedback method in a kind of multipoint cooperative transmission system according to claim 4, is characterized in that, in step 1, the k position PMI information of pending Linear block coding is that the information that need to transmit according to real system determines, by this 2 ^keach k position PMI sequence mapping of individual k position PMI sequence is encoded into the PMI of L position _ithe best mapping code rule of ' sequence is, makes to need after the k position PMI information sequence of the transmission Linear block coding described in carry out step 1, then carries out the resulting code word of network code with SRS

should there is good autocorrelation and cross correlation.

Images